Drowsiness detection system with alarm notification using haar cascade algorithm

ABSTRACT: Road safety is a major concern, and the statistics of fatal crash caused by drowsy driving highlight the need to develop creative approaches that could handle the problem. This research is undertaken with the aim of improving road safety and it builds a drowsiness detection system. The study is constructed to achieve three main goals to be focused on the most crucial issues of the topic. First, there is a focus on developing the drowsiness detection system accuracy through machine learning algorithms such as the Convolutional Neutral Network (CNN) and by using the Haar Cascade pre-trained algorithm. These technologies are conjointly used to get rid of any potential defects in the system, hence enhancing the accuracy, reliability, and effectiveness of the driver drowsiness detection. Secondly, the goal is to adopt a multi-level alarm notification system that will perform the proactive safety tasks other than just reacting to the hazard. The alert system is built so that it gets more sensitive whenever it discovers drowsiness, thus ensuring timely and correct reactions and reducing the possibility of accidents. Finally, similar to before, SMS is used to alert pre-designated persons when users are heavily drowsy. Int involves a proactive communication strategy that allows the entity quick communication of the driver’s condition to external parties like family members, or an emergency contract. Furthermore, the study used a survey-based approach to collect thorough user input on the system’s functions and usability. Notably, the drowsiness detection system was rigously trained and evaluated with the CNN model, producing results for accuracy, specificity, and sensitivity metrics. These metrics demonstrate the system’s performance and robustness in identifying drowsy driving patterns. The research findings show promising results in the achievement of the indicated objectives. The results demonstrated significant improvements in accuracy, the smooth integration of efficient alarm notifications, and the successful implementation of SMS alerts for designated contact persons. These findings highlight the developed system’s potential to greatly reduce the dangers associated with drowsy driving, hence improving overall road safety and user well-being.

5

ABSTRACT: Road safety is a major concern, and the statistics of fatal crash caused by drowsy driving highlight the need to develop creative approaches that could handle the problem. This research is undertaken with the aim of improving road safety and it builds a drowsiness detection system. The study is constructed to achieve three main goals to be focused on the most crucial issues of the topic. First, there is a focus on developing the drowsiness detection system accuracy through machine learning algorithms such as the Convolutional Neutral Network (CNN) and by using the Haar Cascade pre-trained algorithm. These technologies are conjointly used to get rid of any potential defects in the system, hence enhancing the accuracy, reliability, and effectiveness of the driver drowsiness detection. Secondly, the goal is to adopt a multi-level alarm notification system that will perform the proactive safety tasks other than just reacting to the hazard. The alert system is built so that it gets more sensitive whenever it discovers drowsiness, thus ensuring timely and correct reactions and reducing the possibility of accidents. Finally, similar to before, SMS is used to alert pre-designated persons when users are heavily drowsy. Int involves a proactive communication strategy that allows the entity quick communication of the driver's condition to external parties like family members, or an emergency contract. Furthermore, the study used a survey-based approach to collect thorough user input on the system's functions and usability. Notably, the drowsiness detection system was rigously trained and evaluated with the CNN model, producing results for accuracy, specificity, and sensitivity metrics. These metrics demonstrate the system's performance and robustness in identifying drowsy driving patterns. The research findings show promising results in the achievement of the indicated objectives. The results demonstrated significant improvements in accuracy, the smooth integration of efficient alarm notifications, and the successful implementation of SMS alerts for designated contact persons. These findings highlight the developed system's potential to greatly reduce the dangers associated with drowsy driving, hence improving overall road safety and user well-being.

Filipiniana

5